Males make poor meals: a comparison of nutrient extraction during sexual cannibalism and predation
- First Online:
- Cite this article as:
- Wilder, S.M. & Rypstra, A.L. Oecologia (2010) 162: 617. doi:10.1007/s00442-009-1518-3
- 415 Views
Cannibalism is hypothesized to have evolved as a way to obtain a high-quality meal. We examined the extraction of lipid and protein by female wolf spiders, Hogna helluo, during sexual cannibalism of males and predation of crickets. Most food-limited females did not cannibalize males but immediately consumed a size-matched cricket. When consuming male H. helluo and crickets, female H. helluo only consumed 51% of the male body while they consumed 72% of the cricket body. While males had higher protein content in their bodies than crickets and other insects, female H. helluo ingested similar amounts of protein from male H. helluo and crickets. Female H. helluo extracted 47% of the protein present in male H. helluo and 67% of the protein present in crickets. Females were able to extract nearly all of the lipid present in male H. helluo and crickets. However, crickets and other insects had almost 4 times higher lipid content than male H. helluo. The ratio of lipid to protein consumed from crickets appeared more similar to the nutritional requirements of egg production than that of males. Taken together, female hesitancy to engage in cannibalism, low extraction of nutrients from males and a low ratio of lipid to protein in the food extracted from males suggest that males may be poor-quality prey items compared to common insects such as crickets.
KeywordsSexual cannibalismNutritionLipidProteinWolf spider
Cannibalism, the consumption of a conspecific, is a behavior that has long intrigued biologists (Howard 1886; Prete and Wolfe 1992) yet remains poorly understood (Elgar and Crespi 1992; Wise 2006; Wilder et al. 2009a). Cannibalism occurs in a wide range of animals including both vertebrates and invertebrates and between many life history stages (e.g., matricide, filial cannibalism, egg cannibalism, juvenile cannibalism and sexual cannibalism; Elgar and Crespi 1992). In some species, cannibalism occurs relatively frequently and has substantial effects on behavior, morphology and population dynamics (Maxwell 1999; Rudolf 2007; Wilder et al. 2009a). A more thorough understanding of cannibalism can aid in understanding the role of this behavior in evolution, population regulation and food web dynamics (Polis et al. 1989; Wise 2006; Rudolf 2007; 2008).
Several hypotheses have been proposed to explain the evolution of cannibalism including: nutritional benefits to the consumer, reduction of competition, mistaken identity and syndromes of aggression (Elgar and Crespi 1992; Elgar and Schneider 2004; Johnson and Sih 2005; Wise 2006). Of these hypotheses, the idea that cannibalism is a nutritional decision that benefits the consumer has received the most attention in the literature (Elgar and Crespi 1992; Barry et al. 2009). Some authors have also suggested that conspecifics may be high-quality prey since the organism being consumed has the same, or at least very similar, concentrations of nutrients in its body as the consumer (Elgar and Crespi 1992; Pfennig 2000; Denno and Fagan 2003). However, there are conflicting data on the nutritional benefits of cannibalism. Some studies have documented higher growth rates and fecundity of cannibals compared to non-cannibals (e.g., Meffe and Crump 1987; Crump 1990; Snyder et al. 2000) while other studies have failed to detect a difference or have documented a negative effect of cannibalism on growth and reproduction (e.g., Yasuda and Ohnuma 1999; Schausberger and Croft 2000; Mayntz and Toft 2006). Nutritional benefits to the consumer are especially rare following sexual cannibalism, the consumption of a male by a female before, during or after mating (Elgar and Schneider 2004). Of 12 studies that tested if females gain a fecundity benefit of sexual cannibalism, only three found a significant benefit of consuming the male (Birkhead et al. 1988; Rabaneda-Bueno et al. 2008; Barry et al. 2009) while the remaining nine failed to detect a fecundity benefit of cannibalism to females (Andrade 1996; Spence et al. 1996; Arnqvist and Henriksson 1997; Fahey and Elgar 1997; Maxwell 2000; Elgar et al. 2000; Johnson 2001; Stalhandske 2001; Fromhage et al. 2003).
There are several critiques on the idea that cannibalism provides a high-quality diet. First, cannibalism may not provide nutrients in the same ratio as they are present in the consumer’s body because the ratio of nutrients present in individuals of the same species can vary due to developmental state, sex and feeding level (Wilder et al. 2009b). Second, the idea that individuals should consume food with a nutritional composition identical to their own body composition has been challenged (Raubenheimer et al. 2007). To create tissue with a given concentration of energy and nutrients, consumers must ingest food with that amount of energy and nutrients plus extra energy to fuel the metabolic processes needed to obtain and assimilate the food (Raubenheimer et al. 2007). This argument suggests that the ideal diet for a consumer has a higher ratio of energy to nutrients than their own body, not a ratio equal to their own body. In addition, a functional approach to nutrition suggests that individuals should consume food to accommodate their needs, which may or may not be similar to their current body composition. For example, while growing juveniles need to build more somatic tissue in their bodies, adult insects and spiders do not grow in size but rather use food resources for maintenance metabolism and for reproduction (Rickers et al. 2006). Specifically, adult female insects and spiders likely need to consume prey with a high lipid content to accommodate the high lipid content of their eggs and the energy needed for metabolic processes involved in oogenesis (Salomon et al. 2008). This new view of nutritional requirements suggests that the reason so few studies have uncovered a fecundity benefit of sexual cannibalism is that males are not high-quality prey items for females, especially if males have lower lipid content than females (e.g., Wilder et al. 2009b) and egg sac production requires a large lipid investment.
The purpose of this study was to compare the relative nutritional quality of sexual cannibalism and predation. Since there is no significant fecundity benefit of sexual cannibalism in most insects and spiders (Elgar and Schneider 2004; Wilder et al. 2009a), we hypothesized that males were poor-quality food items relative to other insect prey. We first compared the lipid and protein content of fully intact male wolf spiders, Hogna helluo, and crickets, Acheta domesticus, to a wide range of insect prey items. We then allowed female spiders to consume either a male spider or a similarly sized cricket and measured the mass of lipid and protein extracted from each and the extraction efficiency (i.e., the proportion of available lipid or protein that was ingested) by the female spider. Finally, we compared the ratio of lipid to protein present in the egg sacs of female H. helluo with that of the ingested biomass from male H. helluo and crickets to test which type of prey more closely matched the nutritional requirements of egg production.
Materials and methods
The wolf spider, Hogna helluo (Araneae: Lycosidae), is an ideal species with which to test the nutritional benefits of sexual cannibalism. After reaching adulthood, female wolf spiders stop growing in size (e.g., carapace width) and, instead, invest consumed energy and nutrients into producing eggs (Rickers et al. 2006), which contain a large amount of lipid and require metabolically expensive oogenesis. Female H. helluo engage in sexual cannibalism in approximately one-third of mating encounters (Wilder and Rypstra 2008a, b, c). Female H. helluo consume prey items using extraoral digestion (Cohen 1995), during which enzymes are injected into prey, consumable components are ingested and unconsumed parts are discarded. Extraoral digestion makes it possible to examine which nutritional components of prey (e.g., lipid, protein and other compounds) female spiders can ingest and which parts are unconsumed and discarded from prey. Crickets, Acheta domesticus (Orthoptera: Gryllidae), were chosen as representative insect prey because orthopterans are common prey items for wolf spiders in the field (Nentwig 1986; Sokol-Hessner and Schmitz 2002) and because the nutritional composition of A. domesticus is similar to that of 75 species of insects from eight orders (see Results).
Female H. helluo were raised to adulthood in the laboratory at Miami University. These females were the offspring of females collected from the field at the Miami University Ecology Research Center (Oxford, Butler County, Ohio) in 2006. Upon hatching and dispersal from the mother, H. helluo were placed in plastic containers (8 cm diameter × 5 cm high) with a 1-cm substrate of peat moss and soil, a slice of potato and an active culture of the collembolan Sinella curviseta (Collembola: Entomobryidae). Spiders were maintained in an environmental chamber at 25°C and a 13:11-h light:dark cycle. After 3–4 weeks, spiderlings were switched to a diet of two appropriately sized juvenile A. domesticus once or twice per week. Individuals were transferred to larger containers (11 cm diameter × 8 cm high) upon reaching approximately 1 cm body length.
Sexual cannibalism is a relatively infrequent event in H. helluo (i.e., 35% of mating encounters; Wilder and Rypstra 2008a) compared to some other spiders (i.e., 50–100% of mating encounters; Andrade 1996; Knoflach and Van Harten 2001; Schneider and Elgar 2001). Hence, in order to collect sufficient sample sizes of males consumed by females, we collected data from cannibalism events that occurred during two experiments that were designed to examine mating behavior of H. helluo (Wilder and Rypstra 2008c, d). The data used in this experiment were collected after the end of data collection for these two behavioral experiments and do not overlap with the data from the published experiments (Wilder and Rypstra 2008c, d). All females used in both of these experiments were virgins when they were placed into the experimental arenas and were only used once. There were no differences between the two experiments in the mass of the male remaining following feeding (t11.6 = 0.69, P = 0.50) or the mass of lipid (t6.3 = 1.26, P = 0.25), protein (t7 = 1.51, P = 0.17) or other nutrients (t16.4 = 0.17, P = 0.86) extracted from the male body by the female; hence, males from these two experiments were pooled for analysis.
All male spiders were maintained on a diet of two juvenile crickets (ca. 150 mg) per week, which was sometimes more than males could consume. Females (n = 15) from one of the behavioral experiments (Wilder and Rypstra 2008c) were fed five crickets in the week before the experiment and two crickets (ca. 200 mg) in the 2 weeks prior to that (i.e., seven crickets over 3 weeks). Females (n = 27) from the other behavioral experiment (Wilder and Rypstra 2008d) were not fed the week before the trial but were fed three crickets (ca. 200 mg) in the 2 weeks prior to that. There was no significant difference in body condition (mass/carapace width) for females from the two experiments (t40 = 0.6, P = 0.55). Adult female H. helluo are typically maintained on four crickets per week; thus, both of these feeding regimes are food limited for female H. helluo.
In all trials, a female H. helluo was randomly paired with a male and allowed 30–90 min to mate. See (Wilder and Rypstra 2008c, d) for details of mating trials. Trials in which sexual cannibalism occurred were used to test nutrient extraction from male spiders. Whenever one female engaged in sexual cannibalism, the next trial was used to examine nutrient extraction from crickets. If the female did not kill the male, we removed the male and replaced him with a cricket in order to determine if the female was motivated to feed. The cricket fed to the female was matched as closely as possible in mass to the male used in the trial. Female spiders were allowed to feed on male spiders and crickets until they terminated feeding and discarded the prey remains (within 24 h). We also selected 17 male H. helluo and 19 crickets that were approximately 150–200 mg live mass from the laboratory pool of animals to examine the lipid and protein content of fully intact prey.
Nutritional content of eggs
We collected egg sacs from seven adult female H. helluo collected from the Miami University Ecology Research Center. We measured the protein and lipid content present in egg sacs and calculated the ratio of lipid to protein present as the mass of lipid (mg) divided by the mass of protein (mg). Similarly, we calculated the ratio of lipid to protein of the ingested biomass of male H. helluo and crickets as the mass of lipid extracted from each of the prey types by female H. helluo divided by the mass of protein extracted.
We used the ratio of lipid to protein of H. helluo eggs to evaluate the relative quality of ingested food from crickets and male spiders. Female spiders need to consume food that has the same ratio of lipid to protein that is present in the egg sac to allocate those resources to eggs in the appropriate ratio plus an extra amount of lipid to fuel oogenesis (i.e., transforming lipids and proteins into the appropriate types of molecules found in eggs and incorporating them into the eggs). This means that the ideal diet for a female spider would have a ratio of lipid to protein that is higher than that of the eggs. While lipids and carbohydrates are the primary compounds used as metabolic fuel, invertebrates do have the ability to catabolize amino acids to gain energy (Paine 1971, Casas et al. 2005). However, on a per mass basis, the energetic content of amino acids is approximately half that of lipids (Paine 1971). Hence, female spiders could produce eggs on a diet with a lipid to protein ratio lower than their own eggs but they would need to consume a very large amount of protein to compensate for the lack of high-energy lipids and this is unlikely in nature given frequent food limitation of wolf spiders (Wise 1993, 2006). For example, Nyffeler and Benz (1988) found that some wolf spiders may only capture one prey item per day on average. We compared the ratio of lipid to protein in female egg sacs, male H. helluo and crickets using a one-factor ANOVA.
Prey remains or whole prey items were placed in 2-ml centrifuge tubes and stored in a freezer at −20°C until they were analyzed. Prior to analysis, samples were dried at 60°C for 48 h and then weighed to the nearest 0.001 mg. Lipid content was calculated gravimetrically using chloroform extraction. Briefly, dried samples were soaked in chloroform for 24 h after which the chloroform was removed. Each sample experienced three soaking periods and then was re-weighed to calculate the difference in mass before and after lipids were extracted by the chloroform. We packaged 2–3 mg of each sample in a tin capsule and analyzed these samples in a CHN analyzer to measure nitrogen content.
We measured lipid content using chloroform extraction. We estimated protein content by measuring nitrogen content and converting to protein using the standard conversion factor of 6.25 (Horowitz 2002; Mayntz et al. 2005). All calculations were done using nitrogen content and were then converted to protein content directly before statistical analysis. Hence, all P-values for protein also correspond to nitrogen analyses and all measures of protein can be converted to measures of nitrogen by dividing by the conversion factor of 6.25 (Horowitz 2002). The use of protein content as a measure instead of nitrogen allows our data to be compared to other published studies and allowed us to calculate our “other nutrient” category (Wilder et al. 2009b). Data on the protein and lipid content of whole male H. helluo and crickets were compared to published data on the protein and lipid content of 75 species of insects from eight orders using separate one-factor ANOVA for protein and lipid (Bernard and Allen 1997; Ramos-Elorduy et al. 1997).
Since measuring lipids and proteins is destructive, we were not able to measure lipid and protein content before and after a spider fed on a prey item. Instead, we used control data from the whole prey items to calculate a calibration curve. Calibration equations relating the live mass (mg) of control prey to their dry mass and lipid-free lean mass (mg) were created using data on control male H. helluo (dry mass = 0.263 × wet mass + 0.408, n = 17, r2 = 0.88; lean mass = 0.23 × wet mass + 3.057, n = 17, r2 = 0.91) and control crickets (dry mass = 0.397 × wet mass−21.076, n = 19, r2 = 0.77; lean mass = 0.192 × wet mass + 3.121, n = 19, r2 = 0.71). The high r2-values of these regressions indicated that the wet mass of the prey of a given treatment had strong power to predict the nutritional content of that prey item. We then used the live mass of prey items with the calibration equations (calculated from measurements of the content of whole prey) to estimate the content of protein and lipid in prey before it was consumed by a spider.
We calculated the amount of lipid, protein and other nutrients extracted from prey items as the difference between the estimated content of the prey before consumption (using the calibration equation) and the actual measurement of the content of the prey remains. We refer to the consumed part of the prey that was not lipid or protein (i.e., carbohydrates, nucleic acids, vitamins and minerals) as “other nutrients”. The unconsumed category was the mass of prey item remaining after a female spider terminated feeding.
Given that the content of lipid, protein, other nutrients and unconsumed parts of prey items may be interdependent, we first analyzed the data using multivariate ANOVA. Since there was a significant multivariate difference, we then conducted t-tests to examine which of the nutritional measures (i.e., lipid, protein, other nutrients or unconsumed parts) differed between treatment groups.
We also examined extraction efficiency or the proportion of available lipid or protein that was actually extracted from male H. helluo and crickets. To determine this, we compared the mass of protein or lipid extracted from prey with the mass of protein or lipid predicted to be present in prey in separate analyses for protein and lipid. The main factors for this analysis were prey species (male H. helluo vs. crickets) and nutritional status (total amount present in prey vs. amount that was extracted by female H. helluo). The interaction term for the two-factor ANOVA was used to evaluate if extraction efficiency (i.e., the amount of nutritional component consumed vs. the total amount present) differed during consumption of male H. helluo and crickets.
All female H. helluo in this experiment were maintained on food-limited diets and 38% (22 of 58) of the females engaged in sexual cannibalism. Of the females that did not engage in sexual cannibalism, 20 were immediately provided with a cricket after the male H. helluo was removed from the arena and 95% (19 of 20) of them immediately consumed the cricket.
Here we show that male spiders are poor-quality prey for reproducing female spiders compared with other common insect prey. These results are contrary to the long-held hypothesis that phylogenetically similar prey should be of higher quality (Pfennig 2000). Most food-limited female spiders that were confined to an arena with a male spider did not engage in sexual cannibalism. Yet, after the male was removed and a size-matched cricket was added to the arena, these females immediately consumed the cricket, which suggests some reluctance of females to engage in cannibalism. When females did cannibalize males, they only consumed 51% of the male body compared to 72% of the cricket body, indicating that females consume more overall food from crickets than males. In addition, the ratio of lipid to protein that female spiders extract from crickets appears more consistent with what females need for reproduction (i.e., to invest in eggs and fuel oogenesis) than the ratio of lipid to protein extracted from male spiders. Taken together, female reluctance to engage in cannibalism, less overall mass extracted from males and a lower ratio of lipid to protein in the food extracted from males suggest that males may be lower quality prey items compared to common insect prey such as crickets.
Some authors have suggested that predatory arthropods in general are limited by nitrogen (White 1978; Fagan et al. 2002; Denno and Fagan 2003; Fagan and Denno 2004; Matsumura et al. 2004). Nitrogen in the form of protein is likely important for the growth of juvenile arthropods (Mayntz and Toft 2001; Hunt et al. 2004). However, a number of studies have provided evidence that energy (e.g., carbohydrates and lipid) can limit the abundance, activity, longevity and reproduction of predatory arthropods (Helms and Vinson 2002; Wackers et al. 2005; Salomon et al. 2008). For example, in a study of the social spider Stegodyphus dumicola, colonies supplemented with high lipid prey had a higher proportion of breeding females and females that were larger (i.e., prosoma size) and heavier than colonies provided with high protein prey (Salomon et al. 2008). In addition, our data also suggest that lipid, and not protein, may limit egg production of female H. helluo. Female H. helluo can ingest the same mass of protein present in an egg sac (19.8 ± 2.1 mg) from a single cricket (18.6 ± 0.8 mg) or male H. helluo (16.2 ± 1.8; F2,42 = 1.18, P = 0.32). However, female H. helluo cannot produce an egg sac after consuming a single prey item. When female H. helluo are fed one cricket every 3–4 days in the laboratory, they produce egg sacs after 85.2 ± 11.3 days, which suggests that time and energy (e.g., lipids to fuel oogenesis and maintenance metabolism) limit egg production (S. M. Wilder and A. L. Rystra, unpublished data). While few studies have tested the relative effects of lipid and protein on reproduction of arthropods, the importance of lipids and other sources of energy (e.g., carbohydrates, catabolism of amino acids) for reproducing adults should not be surprising given the high costs of reproduction (Williams 1966; Calow 1979; Harshman and Zera 2007). Future studies are also needed to test if other types of nutrients (e.g., vitamins and minerals) are limited in prey and the role of these nutrients for growth and reproduction.
Female H. helluo consumed nearly the entire lipid present in both male H. helluo and crickets, suggesting lipid content of whole prey was a useful measure of lipid available to female spiders. In another study, three different species of adult female wolf spiders also extracted over 90% of the available lipid in a range of prey items (Wilder et al. 2009b). The fact that adult female spiders extract nearly all available lipid in prey suggests that lipids may be a limiting nutrient for adult females. However, the same was not true for protein content. While the whole bodies of male H. helluo had significantly higher concentrations of protein than crickets and other insect prey, female H. helluo extracted similar quantities of protein from male H. helluo and crickets. Female H. helluo extracted 67% of the protein present in crickets while they only extracted 47% of the protein present in male spiders. Similarly, in another study, adult female spiders only extracted 52–80% of the available protein in a range of insect prey (Wilder et al. 2009b). These results highlight the need to measure prey quality by examining what nutritional components predators actually ingest from prey instead of just measuring whole prey items because the nutrients consumed may not be proportional to what was present in whole prey. A difference in the nutritional content of whole prey and what is actually ingested by predators likely occurs for many other predators that selectively consume parts of prey including mammalian carnivores such as felids and canids and avian predators such as falcons, eagles and hawks (Gende et al. 2001).
Sexual cannibalism is a tradeoff between the nutritional benefits of consuming the male body by the female and the costs of consuming a conspecific. Our results suggest that there is some inhibition of cannibalism in hungry female H. helluo. Nearly all of the females that did not engage in cannibalism immediately consumed a cricket similar in size to the male that had just been removed. One cost of sexual cannibalism could be the risk of disease transmission (Pfennig 2000). Sexual cannibalism could also be costly if there is a risk of the male retaliating and damaging the female during a cannibalism attempt, which may explain why females are more likely to cannibalize smaller males (Wilder and Rypstra 2008a, c). In addition to the potentially high cost of cannibalism, our results suggest that the nutritional benefit gained from cannibalism may be relatively low given the low lipid content of male spiders relative to other common insect prey and the need for lipid to invest in eggs and fuel oogenesis (Salomon et al. 2008). The high costs and low benefits of sexual cannibalism may explain why this behavior is not more common and why hungry females avoid attacking and cannibalizing males despite the fact that males are one of the few prey items that actually seek out and approach females.
The prevailing view in many studies of sexual cannibalism is that males are high-quality prey items and that food-limited females should be highly motivated to engage in sexual cannibalism (Elgar and Crespi 1992, Elgar and Schneider 2004). However, our results suggest that food-limited females are reluctant to engage in sexual cannibalism, even when the male is confined in a small arena and could easily be captured. Instead, females may be waiting for insect prey. In our study crickets provided females with a greater abundance of food compared to similarly sized male spiders and food with a ratio of lipid to protein more consistent with what females may need for egg production. The relatively low nutritional quality of male spiders compared to other insect prey may help to explain why sexual cannibalism occurs in relatively few species and why it occurs relatively infrequently in other species (Elgar and Crespi 1992; Wilder and Rypstra 2008a; Wilder et al. 2009a).
We thank J. Moya-Larano and D. H. Wise for comments on a previous draft of this manuscript. We also thank members of the Miami University Spider Lab for collecting and maintaining spiders used in these experiments. Funding was provided by the Department of Zoology at Miami University and by National Science Foundation grant DBI-0216947 to A. L. R.